# Self-supervised learning for classifying paranasal anomalies in the maxillary sinus

## TL;DR

This paper introduces a self-supervised learning method to classify paranasal anomalies in maxillary sinuses using 3D convolutional networks, achieving strong performance with limited labeled data.

## Contribution

A novel self-supervised learning framework for paranasal anomaly classification in maxillary sinuses, leveraging 3D autoencoders and CNNs.

## Key findings

- The proposed SSL method outperforms existing techniques like BYOL and SimSiam in classifying paranasal anomalies.
- With only 10% of labeled data, the method achieves an AUPRC of 0.79, surpassing other SSL approaches.
- The SSL approach effectively localizes anomalies, improving classification of normal vs. anomalous maxillary sinuses.

## Abstract

Paranasal anomalies, frequently identified in routine radiological screenings, exhibit diverse morphological characteristics. Due to the diversity of anomalies, supervised learning methods require large labelled dataset exhibiting diverse anomaly morphology. Self-supervised learning (SSL) can be used to learn representations from unlabelled data. However, there are no SSL methods designed for the downstream task of classifying paranasal anomalies in the maxillary sinus (MS).

Our approach uses a 3D convolutional autoencoder (CAE) trained in an unsupervised anomaly detection (UAD) framework. Initially, we train the 3D CAE to reduce reconstruction errors when reconstructing normal maxillary sinus (MS) image. Then, this CAE is applied to an unlabelled dataset to generate coarse anomaly locations by creating residual MS images. Following this, a 3D convolutional neural network (CNN) reconstructs these residual images, which forms our SSL task. Lastly, we fine-tune the encoder part of the 3D CNN on a labelled dataset of normal and anomalous MS images.

The proposed SSL technique exhibits superior performance compared to existing generic self-supervised methods, especially in scenarios with limited annotated data. When trained on just 10% of the annotated dataset, our method achieves an area under the precision-recall curve (AUPRC) of 0.79 for the downstream classification task. This performance surpasses other methods, with BYOL attaining an AUPRC of 0.75, SimSiam at 0.74, SimCLR at 0.73 and masked autoencoding using SparK at 0.75.

A self-supervised learning approach that inherently focuses on localizing paranasal anomalies proves to be advantageous, particularly when the subsequent task involves differentiating normal from anomalous maxillary sinuses. Access our code at https://github.com/mtec-tuhh/self-supervised-paranasal-anomaly.

The online version contains supplementary material available at 10.1007/s11548-024-03172-5.